Propozycja hybrydowej metody grupowania opartej na sieciach samouczących
A proposal of hybrid clustering method based on self-learning networks
Author(s): Kamila Migdał-NajmanSubject(s): Economy
Published by: Wydawnictwo Uniwersytetu Ekonomicznego we Wrocławiu
Keywords: Self Organizing Map (SOM); Growing Neural Gas (GNG); validity indexes; preferences research; segmentation
Summary/Abstract: In the article a hybrid clustering method based on a self-learning neural networks, SOM and GNG, is presented. The author verified the potential of the proposed hybrid clustering method such as: network SOM + k-means method. The proposed approach is verified on the example of behavioral research and communication preferences of the inhabitants of Gdynia in 2010.
Journal: Prace Naukowe Uniwersytetu Ekonomicznego we Wrocławiu
- Issue Year: 2012
- Issue No: 242
- Page Range: 342-351
- Page Count: 10
- Language: Polish